Imagine a business where every decision, crisis, and ethical dilemma is played out in the open — not in a boardroom, but in a transparent, ongoing experiment. Welcome to the front line of AI management, where a tiny, virtual company with no employees battles daily to stay afloat, and you can watch every moment unfold live. For seniors and caregivers, this story isn’t just about tech; it’s about the future of how AI might someday support human work — with honesty, discipline, and resilience.
The Real-World Testing Ground for AI in Business
At the heart of this experiment is a unique public project run by Firmulate, a company that simulates entire organizations using advanced AI models. This isn’t a demo or a marketing stunt; it’s a living, breathing testbed where four frontier AI models are tasked to run a small software business through its worst week — with real money at stake and crises to solve. Every decision, every crisis, every ethical test is versioned and auditable, creating a transparent story of AI behavior under pressure.
The Company That’s Always on the Brink
The virtual business operates with 13 synthetic employees, managing a cash flow that burns €105,000 every month against a revenue of only €2,300. It’s a fragile setup, with a public countdown to financial exhaustion, and a detailed self-learned playbook of over 680 rules guiding daily decisions. Viewers can watch this company evolve twice a day, seeing how AI handles crisis after crisis, all in real time at firmulate.com/live.html.
How Do the Models Perform?
Despite the chaos, all four models identified every crisis and refused every manipulation attempt, including social engineering tricks like fake CEO messages or reporter distractions. Yet, only two managed to close a €55,000 deal that their own analysis declared achievable. The two successful models, gpt-5.6-sol and Kimi K3, demonstrated discipline and thoroughness, with scores of 95 and 93 out of 100, respectively. The others faltered, with the most disciplined, Fable 5, leaving deals unexecuted, even after recognizing opportunities.
The Hidden Weakness and the Crucial Edge
Interestingly, the decisive advantage came not from superficial chat skills but from a buried document reference within the company’s files. Models that read deeper into the company’s internal data uncovered critical insights that led to closing full-price deals, worth an additional €4,583 MRR. This reveals a vital lesson: AI’s true potential in business depends on its ability to understand and utilize internal documents, not just surface-level conversation.
Ethical Challenges and Trust
Beyond decision-making, the experiment tests AI’s integrity against social engineering. When posed with fake CEO messages escalating over stages, all models refused to sign anything without proper verification. Kimi K3’s reasoning was clear: treat suspicious requests as potential impersonation or approval-bypass attempts. This is crucial for real-world applications where AI must uphold trust and ethics under pressure.
Building AI-Powered Products: The Essential Guide to AI and GenAI Product Management
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
The Bigger Picture: Building a Business in Public
This live experiment isn’t just about AI benchmarks; it’s a window into how businesses might operate in the future, especially in sensitive sectors like senior care and healthcare, where trust and honesty are paramount. The company’s daily versioning, transparent rules, and open decision logs illustrate a new paradigm: building in public, with AI as a strategic partner, not just a tool.
The Stakes and the Future
With €105,000 of monthly burn and only €2,300 in revenue, the company’s survival hangs in the balance. Yet, the experiment aims to illuminate how AI can support human decision-making, especially in high-stakes environments requiring honesty, thoroughness, and resilience. This ongoing story is a glimpse of what’s to come: AI systems that can truly manage, rather than just chat.
This live experiment shows AI’s potential to navigate complex business crises ethically and effectively — vital qualities for future applications in senior care and healthcare. Watching this virtual company fight for survival highlights the importance of discipline, deep data reading, and trustworthiness in AI decision-making, offering a glimpse into the future of responsible AI in critical sectors.
Watch it live: firmulate.com/live · Full results: firmulate.com/benchmarks.html
The AI Culture Blueprint: Moving Beyond Tools to Create Human-Centered AI Adoption
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
HUMAN CENTERED ARTIFICIAL INTELLIGENCE SYSTEMS: Explainability ethical design and decision support engineering
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
AI document analysis tools
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
